fix: Adjust reconcile duration histogram buckets for millisecond-level precision

[forkline-bot]

Summary

• Adjusted histogram bucket boundaries to capture millisecond-level reconcile durations
• Previous buckets (0.1, 0.5, 1.0, 5.0, 10.0 seconds) were too coarse for actual reconcile times averaging ~0.9ms
• New buckets provide better granularity: 0.001s, 0.005s, 0.01s, 0.025s, 0.05s, 0.1s, 0.25s, 0.5s, 1.0s, 2.5s, 5.0s, 10.0s

Problem

The current histogram buckets were not useful for monitoring because all reconcile operations completed in less than 0.1 seconds (the smallest bucket). This made it impossible to observe performance variations in the typical reconcile duration range.

Example metrics showed:

robotlb_reconcile_duration_seconds_bucket{controller="robotlb",le="0.1"} 111
robotlb_reconcile_duration_seconds_bucket{controller="robotlb",le="0.5"} 111
robotlb_reconcile_duration_seconds_bucket{controller="robotlb",le="1"} 111
...
robotlb_reconcile_duration_seconds_sum{controller="robotlb"} 0.00090932

All values were in the first bucket, with an average of ~0.9ms.

Solution

New bucket boundaries in seconds:

• Sub-millisecond: 0.001 (1ms)
• Milliseconds: 0.005 (5ms), 0.01 (10ms), 0.025 (25ms), 0.05 (50ms)
• Hundreds of milliseconds: 0.1 (100ms), 0.25 (250ms), 0.5 (500ms)
• Seconds: 1.0, 2.5, 5.0, 10.0

This captures the typical sub-10ms reconcile times while still tracking potential outliers in the second range.

Resolves: #54

[pando85]

@forkline we don't need so many values. Reduce the number of ranges in the bucket, do it in a useful way, select the most relevant for the operations

[forkline-bot]

Done. Reduced histogram buckets from 12 to 5 essential values:

• 0.001s (1ms): Typical reconcile time (~0.9ms average)
• 0.01s (10ms): Slightly slower operations (10x typical)
• 0.1s (100ms): Slow reconciles (100x typical)
• 1.0s: Significant outliers (1000x typical)
• 10.0s: Extreme cases

This provides a clean logarithmic distribution that captures the full performance range while minimizing cardinality.

[forkline-bot]

✅ CI Fix Applied

I've pushed a fix for the CI failure. The changes should trigger a new CI run.

Commit:

Waiting for CI to re-run...